Home General Libraries GridRPC Documents Tests Contact Misc
Mathematical libraries

One of the aims of the our project as well as the NetSolve and Ninf projects is to provide the scientific users with easy, uniform and effective access to the numerous mathematical libraries installed in educational computing centres.

Mathematical libraries provide the implementations of well-known and often-used computational operations, such as linear algebra operations (matrix operations, linear equation systems solvers), Fourier transformates, random numbers generation, statistical analysis etc. These operations are grouped in the functionality sets, such as BLAS (Basic Linear Algebra Subprograms) and LAPACK (Linear Algebra Package).

There are many implementations of individual mathematical functionality sets. The platform- and architecture-optimised versions are able to exploit the power of underlying computing systems (SMPs, MPPs, clusters etc.) effectively. The programmer gains access to the power of well-designed numerical algorithms and well-equipped computing nodes without any specific (computing science-related) knowledge simply by linking his/her application with a mathematical library.

Several mathematical packages that we exploit in Polish computing centres along with the information on which functionality sets they implement are presented in the table.



Functionality BLAS Sparse BLAS BLAS subset LAPACK Linpack Eispack FFT
Complib (SGI)    
ESSL (IBM)    
MLIB (HP)
Math Kernel Library (Intel)      
LIBSCI (CRAY)  
NAG Fortran Library
Netlib